Research Article
Feasibility Pump Algorithm for Sparse Representation under Laplacian Noise
Table 1
Mean relative errors (
11) for SFP (top in each cell), SFPreg (middle), and RLAD (bottom) for matrix conditioning of 1000.
| | 5 | 7 | 9 | 11 |
| SNR = 10 | 0.262 | 0.379 | 0.427 | 0.471 | | 0.255 | 0.255 | 0.248 | 0.241 | | 0.447 | 0.450 | 0.500 | 0.471 |
| SNR = 20 | 0.084 | 0.143 | 0.378 | 0.338 | | 0.084 | 0.084 | 0.082 | 0.082 | | 0.201 | 0.276 | 0.410 | 0.415 |
| SNR = 30 | 0.027 | 0.027 | 0.095 | 0.157 | | 0.027 | 0.026 | 0.025 | 0.025 | | 0.068 | 0.231 | 0.270 | 0.464 |
| SNR = | | | | | | | | | | | | | | |
|
|